IOP Publishing Journal of Physics: Conference Series 60 (2007) 58–65 doi:10.1088/1742-6596/60/1/010 Proceedings of the TeV Particle Astrophysics II Workshop Search for Weakly Interacting Massive Particles with CDMS and XENON Elena Aprile1,2, Laura Baudis3,4, Blas Cabrera5,6 1 Physics Department and Columbia Astrophysics Laboratory, Columbia University, New York, NY 10027, USA 3 Department of Physics, RWTH Aachen University, Aachen, 52074, Germany 5 Physics Department and KIPAC, Stanford University, Stanford, CA 94305, USA 1 3 E-mail: [email protected], [email protected], 5 [email protected] Abstract. The Cryogenic Dark Matter Search (CDMS) and XENON experiments aim to directly detect dark matter in the form of weakly interacting massive particles (WIMPs) via their elastic scattering on the target nuclei. The experiments use different techniques to suppress background event rates to the minimum, and at the same time, to achieve a high WIMP detection rate. The operation of cryogenic Ge and Si crystals of the CDMS-II experiment in the Soudan mine has yielded the most stringent spin-independent WIMP-nucleon cross- section (∼ 10−43 cm2) at a WIMP mass of 60 GeV/c2. The two-phase xenon detector of the XENON10 experiment is currently taking data in the Gran Sasso underground lab and promising preliminary results were recently reported. Both experiments are expected to increase their WIMP sensitivity by a one order of magnitude in the scheduled science runs for 2007. 1. Introduction Recent observations from high-redshift supernovae [1, 2, 3], cosmic microwave background anisotropy measurements [4, 5], the red-shift galaxy clusters [6], and the Sloan Digital Sky Survey [7] provide growing evidence that the mass in the Universe is dominated by dark matter, which is non-luminous, non-baryonic and could be weakly interacting massive particles (WIMPs) [8]. One attractive candidate for WIMPs is the lightest supersymmetric particle (LSP) from supersymmetry models [9]. A WIMP can deposit a small amount of energy (a few tens of keV) in an elastic scattering with an atomic nucleus [10], with a rate much less than 1 event/kg/day. Figure 1 shows the predicted event rates for a WIMP mass of 100 GeV/c2 and a spin- independent WIMP-nucleon cross-section of 10−43 cm2 for different target materials, according to [11]. It can be seen that the event rate falls quickly at high recoil energy. A low energy- threshold (∼10 keV) is one of the key requirements for direct dark matter search experiments, such as CDMS and XENON, to achieve a good WIMP detectability. While the WIMP detection rate on xenon (the target material for XENON experiment) is suppressed at large recoil energies (we use the nuclear form factor by [12]), it is enhanced at low recoil energies by the large atomic 2 On behalf of the XENON collaboration 4 On behalf of the CDMS and XENON collaborations 6 On behalf of the CMDS collaboration © 2007 IOP Publishing Ltd 58 59 Figure 1. Calculated WIMP detection rate as a function of detector energy-threshold Eth for four different target materials (Xe, Ge, Ar and Si). A WIMP mass of 100 GeV/c2 and WIMP-nucleon cross-section of 10−43 cm2 are used. −1 10 Si (A = 28) Ar (A = 40) −2 10 Ge (A = 73) Xe (A = 131) −3 10 −4 10 Total Rate [evt/kg/day] −5 10 0 10 20 30 40 50 60 70 Eth - Nuclear Recoil Energy [keV] mass of xenon. From figure 1, the event rate in the Xe detector is about 30% higher than for Ge, for an energy threshold of 10 keV. The CDMS experiment uses both Ge detectors for high WIMP sensitivity (×6 better than Si) and Si detectors for added neutron rejection with the same neutron rate as Ge. 2. CDMS-II Experiment The CDMS-II experiment (a detailed description of the experimental apparatus can be found in [13]) is located in one of the two excavated caverns of the Soudan Underground Laboratory, at a depth of 780 m (2090 m.w.e. or meter-water-equivalent). The surface muon flux is reduced at the underground lab by a factor of 5 ×104. An additional active muon veto shield, consisting of 5-cm-thick BC-408 plastic scintillator slabs, is used to tag cosmic ray muons. A passive shield, consisting of 40-cm-thick outer polyethylene, 22.5-cm-thick lead (including 4.5-cm-thick inner ancient lead), and 10-cm-thick inner polyethylene, is enclosed by the muon veto and houses the detectors. The passive shield reduces most of the gamma-ray and neutron background from radioactivity in the surrounding rock or produced by cosmic ray muons interacting with the surrounding rock and shield. To reduce the radon activity in the vicinity of the detectors, medical grade breathing air is used to purge the space inside of the Pb shield continuously. At the core are Z(depth)-sensitive ionization and phonon-mediated (ZIP) detectors, kept at a base temperature below 50 mK. Each ZIP detector is a cylindrical high-purity Ge (250 g) or Si (100 g) crystal. Two concentric ionization electrodes and four independent phonon sensors are photolithographically patterned onto each crystal. A particle interacting in a ZIP detector causes either an electron recoil (gamma rays, electrons, neutron inelastic scattering, etc.) or a nuclear recoil (neutron elastic scattering, WIMPs). The interaction deposits its energy into the crystal through charge excitations (electron-hole pairs) and lattice vibrations (phonons). The charge excitations are collected with electrodes on the two sides of the ZIP detector. The electron-equivalent energy EQ is equal to NQ × ,whereNQ is the number of collected charges and is the average energy needed to produce an electron-ion pair ( ≈ 3eVinGeand3.8eV in Si). The phonon signal is detected by quasiparticle-trap electrothermal-feedback transition- 60 edge sensors (QETs) photolithographically patterned onto one of the crystal faces. Since the drifting of electrons and holes across the crystal also contributes the phonon signal, the recoil energy ER of an event is inferred from both the phonon and ionization signals. EQ is equal to the recoil energy ER for electron recoils. Nuclear recoils produce less electron-ion pairs, which makes EQ less than ER. Thus the ionization yield, defined as y = EQ/ER, is much smaller for nuclear recoils (y ∼ 0.3 for Ge and ∼ 0.25 for Si above 20 keV) than that for electron recoils (y = 1). It provides the technique to reject the electron-recoil events which produce most of the background. Figure 2. Ionization yield versus recoil energy for events in the CDMS-II (top - Ge, bottom - Si) detectors, before the timing cut for removing the surface electron recoil events. The curved lines represent the defined WIMP signal regions with an energy-threshold at 7 keV (dashed vertical lines). After the cut, one event (star, red) remains in the signal region for the Ge data, while no candidate event is seen for the Si data. Figure from [14]. 0.6 0.4 0.2 Ionization Yield Ge 0 0.6 0.4 0.2 Ionization Yield Si 0 5 10 20 30 40 50 60 70 80 90 100 Recoil Energy (keV) The CDMS-II experiment has been operating in the Soudan mine since October 2003 and reported their recent results, from WIMP search data collected through August 2004 with a total of 74.5 live days exposure, yielding a total spectrum-weighted exposures of 34 (12) kg d for the Ge (Si) targets in the 10-100 keV nuclear recoil region [14]. Analysis was performed blindly by masking the events from the candidate region, defined by using calibration data from 133Ba and 252Cf sources and from non-masked WIMP search data. The ZIP detector provides excellent event-by-event discrimination of nuclear recoils from background electron recoils in the detector’s bulk region. However, electron recoils near the detector’s surface suffer from poor ionization collection and leak into the candidate nuclear recoil region. New analysis techniques were developed to remove those surface electron recoils. The phonon pulses from surface recoils are more prompt than those recoils in the detector bulk. A timing parameter, defined as the sum of the time delay of the phonon signal relative to the fast ionization signal and the phonon rise time, was used to reject the surface electron recoils. One candidate event was found after unmasking the Ge WIMP search data (figure 2 top). However, that event occurred in a time period when the detector suffered inefficient ionization collection. It is also consistent with the rate of expected background. The expected number of surface events after the timing cut is 0.4±0.2(stat)±0.2(sys) between 10-100 keV in Ge. The expected neutron background that 61 escapes the muon veto is 0.06 events in Ge. No candidate event was found in the Si data (figure 2 bottom). Based on these data and previous results from Soudan [15], a 90% C.L. upper limit on the spin-independent WIMP-nucleon cross section is 1.6×10−43 cm2 from Ge and 3×10−42 cm2 from Si, for a WIMP mass of 60 GeV/c2. The experiment is now operating five towers (21 Ge and 9 Si ZIP detectors) with more than 4 kg of Ge and 1 kg of Si and is expected to reach a sensitivity down to 2.1 × 10−44 cm2 for spin-independent WIMP-nucleon cross section at 60 GeV/c2 by the end of 2007. In addition, plans are underway for the construction and operation at SNOLAB of the SuperCDMS 25 kg Experiment, which will operate 25 kg of Ge for three years with a target sensitivity of 10−45 cm2 for 60 GeV/c2 WIMPs.